Terminator: Rise of the AI sceptics

As Republican Party reels from the landslide defeat in last week's US midterm elections, the post-mortem results seem clear: It's the war, stupid!

Party officials all the way up to J D Vance acknowledged that the economic distress caused by an ill-considered and badly sold war with Iran turned voters off the ruling party. But Democrat triumph will have impacts not just on US policy in West Asia but also on the country's most important sunrise industry: AI.

Democrats across the country ran, broadly speaking, on an 'anti-AI' platform. The new darling of the party's Left, senator-elect Abdul El-Sayed of Michigan, called for a pause in AI development. Georgia senator-elect Jon Ossoff, standard bearer of the party's centrists, also highlighted in his victory speech a dark vision of a world in which 'the titans of Silicon Valley dig bunkers for themselves and warn that the artificial intelligence they're creating risks human extinction'.


Also Read | Chipmakers warn AI demand is creating bottlenecks in automotive supply chain

Democrat AI-scepticism is on a collision course with a Trump regime that is still gung-ho on the tech. The Trump regime has focused on the need for the US to 'win the AI race'. Beyond flashy PR moves like rebranding AI as 'super intelligence', it has thrown real weight behind the industry.

Trump has argued that that narrow readings of copyright law cannot be permitted to strangle the industry in its crib, and that the pool of human knowledge should be available to train AI models. Perhaps, most importantly, Trump has used presidential executive powers to attempt to ensure that the AI industry does not become enmeshed in a thicket of overlapping, conflicting state-level laws, and is subject only to a single, federal standard.

Here, we are confronted by an apparent paradox: industry does not seem unanimously enthused by the Trump regime's efforts in this direction. Anthropic, for example, has endorsed, and reportedly even lobbied for, tougher state-level safety standards on AI models.

Also Read | Anthropic eyes $1.4 trillion valuation in blockbuster IPO, targets up to $90.4 billion in fundraising

OpenAI, though less vocally 'doom-er' than Anthropic, has also joined in on calls to 'pace the frontier' of AI development. So far, CEOs have obediently lined up at the Trump regime's gala events celebrating 'super intelligence' and smiled through gritted teeth. But they do not appear fundamentally to agree to the 'let it rip' approach.

David Sacks, VC and one-time Trump AI czar, has not minced his words. He has argued openly that Anthropic is pushing tougher regulations, which only entrenched incumbents have financial and administrative resources to understand and comply with. He has dubbed Anthropic's approach a 'sophisticated regulatory capture strategy based on fear-mongering'. He has also pushed back against the call to 'pace the frontier', essentially arguing that OpenAI and Anthropic are free to develop their product as slowly as they like, but should not invite the government to gum up the works for others.

AI safety-ists, however, now have powerful allies in the new Democrat legislative majority. The party is unlikely to have much love for VCs and other members of the 'tech Right' it sees as having betrayed the values of liberal Silicon Valley. Anthropic and OpenAI may now see light at the end of the tunnel: a post-Trump Washington may now see a return of woke sensibilities, and be more trigger-happy about regulating in response to perceived 'harms'.

Certainly, the populist backlash against data centres isn't in the interests of anybody in the AI industry - incumbents, new entrants, Right or Lefz. But on question of AI safety and regulation, AI incumbents will certainly find open ears in the new Democrat majority.

Milton Friedman once argued that the two biggest enemies of the free market were his fellow intellectuals and the business community: the former think they know best and should dictate on questions of economic planning; the latter don't want competition from new entrants.

So far, the Trump regime has stayed largely true to its conservative credentials in staving off regulatory capture of AI. Indeed, Trump has thumbed his nose at AI-sceptics even on the populist Right. But he has now lost valuable allies for his approach in the legislature. He will now have to face down a legislature dominated by the 'party of intellectuals', and an AI business community no longer without allies in positions of governmental authority.